An efficient Machine Learning Techniques for Early Detection of Hearing Loss

Abstract

By 2050, over 700 million people will have severe hearing loss. Audiologists and otolaryngologists are in short
supply in underdeveloped and emerging countries, where a considerable part of the population suffers from
incapacitating hearing loss. Most hearing impairments are untreated for long periods of time due to a scarcity of
specialists. In this study, we present automated hearing impairment diagnosis software based on machine
learning to help audiologists and otolaryngologists consistently and effectively identify and classify hearing
loss.We discuss the architecture, implementation, and performance evaluation of the two-module automated
program for diagnosing hearing impairments: a machine learning model and a module for creating hearing test
data. To train and evaluate the machine learning model, the Data Acquisition Module generates a sizable and
comprehensive dataset. The kind, degree, and arrangement of hearing loss can be accurately predicted by the
model in real time using multiple classes and multi-label classification algorithms that learn from hearing test
data.With a log loss reduction rate of 98.48%, a prediction time of 634 ms, and macro and micro precisions of
100%, our proposed machine learning model shows promise and can help audiologists and otolaryngologists
quickly and accurately classify the type, degree, and configuration of hearing loss.